Sensitivity of radar data on landfall processes of tropical cyclones in the Bay of Bengal

The eastern coastal parts of India, specifically West Bengal, Odisha, Andhra Pradesh, and Tamil Nadu, are always susceptible to intense tropical cyclone (TC) landfalls formed over the Bay of Bengal. It is extremely challenging to accurately forecast the TC structure, intensity, and rainfall, particularly during landfall hours, which have a direct and immense impact on disaster preparedness and mitigation efforts. During landfall the eyewall associated convective processes are complex and puts a constraint in the accurate TC rainfall estimation for the numerical models. This work aims to fill that gap simulating three landfalling TCs, Hudhud (2014), Vardah (2016), and Titli (2018), with a lead time of up to 72 h, by assimilating Doppler Weather Radar (DWR) reflectivity (Rf) observations. Two sets of experiments are performed, i.e., CNTL (without assimilation) and RDA (with assimilation). Results demonstrate a significant improvement in TC intensity (28.9%, 21.4% in minimum central pressure, and 29.6%, 17% in maximum sustained surface wind) for Hudhud and Vardah, respectively, compared to CNTL. However, the improvement on Titli is minimal. Additionally, the RDA has shown an improved rainfall skill score for all cases, particularly over land, i.e., Hudhud (15.5%) and Vardah (53.68%). The hydrometeor distribution near the eyewall reveals a realistic representation in RDA, thereby contributing to improvements in intensity and rainfall forecast. In general, during landfall hours, the radial and tangential wind structures in RDA reveal more organised eyewall convection and are coherent with observations. Furthermore, the horizontal transport term is the dominant contributor to angular momentum influx into the core region and regulating the intensity. Simultaneously, the vertical eddy momentum loss has been reduced in RDA, facilitating the maintenance of the eyewall structure, especially in the case of Hudhud and Titli.